U.S. patent number 6,067,043 [Application Number 09/274,471] was granted by the patent office on 2000-05-23 for pulse compression radar.
This patent grant is currently assigned to Alcatel. Invention is credited to Alain Faure, Noel Suinot.
United States Patent |
6,067,043 |
Faure , et al. |
May 23, 2000 |
Pulse compression radar
Abstract
The invention concerns a method of synthesizing a replica used
in the compression filter of pulse compression radar. The replica
is calculated from the spectrum of a required impulse response. The
required impulse response is preferably obtained from an analytical
function, such as a sinc function or a weighted sinc function, and
a template. It is preferable to use calibration signals of the
instrument to calculate the replica. The invention also applies to
synthetic aperture radar.
Inventors: |
Faure; Alain (Pins Justaret,
FR), Suinot; Noel (Escalquens, FR) |
Assignee: |
Alcatel (Paris,
FR)
|
Family
ID: |
9524362 |
Appl.
No.: |
09/274,471 |
Filed: |
March 23, 1999 |
Foreign Application Priority Data
|
|
|
|
|
Mar 23, 1998 [FR] |
|
|
9803529 |
|
Current U.S.
Class: |
342/174; 342/135;
342/204; 342/202 |
Current CPC
Class: |
G01S
7/4052 (20130101); G01S 13/282 (20130101); G01S
7/4082 (20210501); G01S 13/90 (20130101) |
Current International
Class: |
G01S
7/40 (20060101); G01S 13/00 (20060101); G01S
13/28 (20060101); G01S 13/90 (20060101); G01S
007/40 () |
Field of
Search: |
;342/174,168,169,172,196,202,203,204,189,134,135,145,159,162 |
References Cited
[Referenced By]
U.S. Patent Documents
|
|
|
5329284 |
July 1994 |
Groenenboom et al. |
5568150 |
October 1996 |
Taylor, Jr. et al. |
5708436 |
January 1998 |
Loiz et al. |
5808580 |
September 1998 |
Andrews, Jr. |
|
Foreign Patent Documents
|
|
|
|
|
|
|
0 291 337 A2 |
|
Nov 1988 |
|
EP |
|
2 473 184 |
|
Jul 1981 |
|
FR |
|
42 40 225 A1 |
|
Jun 1994 |
|
DE |
|
Other References
M I. Skolnik, Introduction to Radar Systems, 1981, McGraw-Hill,
Singapore, XP00208524, pp. 426-427. .
N. Vincent et al., Spaceborne Rain radar Mission and Instrument
Analysis, IGARSS 1996, International Geoscience And Remote Sensing
Symposium Remote Sensing For A Sustainable Future, Lincoln, NE, May
28-31, 1996, vol. 1 pp. 279-281, XP000659601. .
N. Vincent et al, "Pulse Compression with -65 DB Sidelobe Level for
a Spaceborne Meteorological Radar", IGARSS 1996, International
Geoscience And Remote Sensing Symposium Remote Sensing For A
Sustainable Future, Lincoln, NE, May 28-31, 1996, vol. 1 pp.
490-492, XP00659613. .
A. Netterstrom, "Using Digital Pre-Distortion to Compensate for
Analog Signal Processing Errors" International Radar Conference,
Arlington, May 7-10, 1990, pp. 243-248, XP000131186, Institute Of
Electrical And Electronics Engineers. .
"Polytime coding as a means of pulse compression", Fielding, J.E.,
Aerospace and Electronic Systems, IEEE Transactions on vol. 35 2,
Apr. 1999, pp. 716-721, 1999. .
"Unfocused SAR using a Next-Generation COoherent Radar Depth
Sounder for measurement of Greenland ice sheet Thickness",
Legarsky, J.; Gogineni, S.P., Geoscience and Remote Sensing
Symposium Proceedings, 1998.IGARSS '98. 1998 IEEE International
vol. 1, 1998. .
"Gradient descent method for designing optimum radar signals",
Skinner, B.J.; Donohoe, J.P.; Ingels, F.M.; Radar Conference,
1995., Record of the IEEE 1995 International, 1995, pp. 67-71,
1995..
|
Primary Examiner: Sotomayor; John B.
Attorney, Agent or Firm: Sughrue, Mion, Zinn, Macpeak &
Seas, PLLC
Claims
We claim:
1. A method of synthesizing a replica used in the compression
filter of pulse compression radar wherein the replica is calculated
from a required impulse response, wherein the required impulse
response is obtained from an analytic function such as a sinc
function or a weighted sinc function and from a template, wherein
calibration signals of the instrument are used to calculate the
replica, and wherein the replica is defined by the following
equation:
in which RI.sub.obj (f) is the spectrum of the required impulse
response and S.sub.cal (f) is the spectrum of a calibration signal,
in particular a calibration pulse.
2. A method according to claim 1, wherein, in the required impulse
response spectrum, an upper frequency band of approximately 10% to
100% of the active band is adopted to obtain the required
resolution.
3. A method according to claim 1, wherein a calibration signal or
pulse is used to synthesize the replica which is measured and/or
estimated and which represents the distortion introduced by the
radar.
4. A method according to claim 3, wherein the replica is
periodically measured and/or estimated in order to adjust it
periodically.
5. A method according to claim 3, wherein measuring the calibration
pulse consists in acquiring a signal that has passed through the
transmit and receive subsystems of the radar.
6. A method according to claim 2, wherein the spectrum of the
required impulse response is obtained by applying a discrete
Fourier transform algorithm to the required impulse response in the
time domain.
7. Application of the method according to claim 1 to coherent radar
over the time to integrate a plurality of pulses.
Description
The invention concerns pulse compression radar.
BACKGROUND OF THE INVENTION
In radar of this type the frequency of the wave transmitted towards
the target varies as a function of time, for example in a linear
fashion. To avoid degrading the distance resolution on reception,
the wave is time compressed. The filter applying this compression
uses a signal which is often called a "replica".
The replica determines the performance of the radar, which is
usually characterized by its impulse response. Impulse response is
the response of the instrument when the signal is transmitted
towards an isolated point target, without interference. Impulse
response in the time domain RI(t) has the value:
where S(t) is the transmitted signal as a function of time, r(t) is
the replica and * is the convolution product.
In the frequency domain:
where S(f)=F(s(t)) is the spectrum of the transmitted signal and
R(f)=F(r(t)) is the spectrum of the replica.
In the conventional case of a matched filter, which maximizes the
signal to noise ratio:
where s*(t) is the complex conjugate-of the signal s(t).
In this case, the impulse response is a sinc function (sin t)/t
having an active central peak and secondary lobes that must be
attenuated or eliminated as they degrade the quality of radar
measurements because they can be interpreted as false targets and
reduce commensurately the energy contained in the active main lobe.
The usual method of attenuating the secondary lobes of the impulse
response is to multiply the replica by a weighting function.
However, weighting widens the main lobe and reduces its amplitude
maximum. Widening (increase in duration) degrades resolution, and
reduction in amplitude degrades signal to noise ratio. To obtain
the required resolution and signal to noise ratio, degradation is
compensated by increasing the transmitted frequency band and
transmitter power. The radar is therefore complex and costly. Also,
the far lobes, also known as Fresnel lobes, are not attenuated
significantly because the level of these lobes is -20.log(BT)+3 dB
away from the maximum of the main lobe, B being the bandwidth of
the transmitted signal and T the duration of the transmitted
pulse.
To attenuate the Fresnel lobes, the transmitted signals depart from
a signal with simple linear variation of frequency. This also
increases the complexity of the radar because it increases the
bandwidth of the signals transmitted and received.
OBJECTS AND SUMMARY OF THE INVENTION
The invention remedies these drawbacks by synthesizing the replica
from the required impulse response.
In other words, the required impulse response (in the time domain)
is fixed a priori, i.e. the attenuation of the secondary lobes, the
level of the Fresnel lobes and the other parameters of the impulse
response are all fixed in accordance with the required performance.
The replica is obtained digitally from the Fourier transform of the
required impulse response in the time domain.
A discrete Fourier transform algorithm is preferably used, such as
a fast Fourier transform (FFT) algorithm.
The Fourier transform of the required impulse response provides a
spectrum that is usually unlimited, i.e. that extends to infinity.
In practice it is therefore necessary to select the frequency band
retained after taking the Fourier transform. It is possible either
to make do with the active band (determined by the required
resolution), or else to include areas outside the active area to
enhance the possibilities of forming an impulse response that meets
the objective. The bandwidth increase lies in the range 10% to
100%, for example. The bandwidth increase must be limited because
it increases the noise band and degrades the signal to noise ratio
commensurately.
The method in accordance with the invention of synthesizing the
replica is not limited to a particular bandwidth for the signal to
be transmitted. It is compatible with conventional transmitted
signals with linear frequency variation and with wider bandwidth
signals corresponding to a non-linear frequency/time law.
Because the replica must be multiplied by the spectrum of the
signal received by the radar and this signal is necessarily
affected by the inevitable distortion introduced by the instrument,
it has been found that to obtain a correct impulse response it is
preferable to allow for such distortion in determining the
replica.
This is why, in the preferred embodiment of the invention, the
instrument is calibrated and the calibration signal is used to
calculate the replica.
To this end, a calibration impulse response is determined, the
spectrum is calculated and its inverse is multiplied by the Fourier
transform (spectrum) of the required impulse response.
Calibration can be performed in various ways. In a first example,
distortion is measured after the transmitted signal has passed
through the receive subsystem. The calibration signal is then
sampled and digitized.
The distortion is measured periodically, for example at intervals
of one minute or each time the radar is started up, in order to
adjust the replica periodically.
In a second example, which can be combined with the first,
calibration is performed by estimation, i.e. by calculation. To
this end the measurements concerning the various units of the radar
or the predicted performance are
combined by calculation to obtain an estimate of the signal in the
form required after passing through all of the radar.
The performance of the radar depends equally on the accuracy and
frequency with which the calibration pulse is calculated, because
any degradation seen by the active signal but not by the
calibration signal cannot be fed into the calculation of the
replica and therefore degrades the impulse response. The
calibration method must therefore be chosen so that non-calibrated
distortion enables the impulse response to satisfy the imposed
conditions, i.e. to meet the specifications.
The invention applies not only to pulse compression radar but also
to radar coherent over a plurality of pulses, such as synthetic
aperture radar (SAR). In the foregoing description and in the
following description the term "pulse compression radar" must be
understood as referring not only to radar of this type but also to
coherent radar in which the received signal is subject to frequency
variation (linear or otherwise) pulse by pulse, as is the case with
the Doppler effect in synthetic aperture radar during passage of
the echo transversely to the antenna lobe as a consequence of
movement of the moving body carrying the radar. In this latter case
the distortion measured by the calibration pulse consists in the
errors occurring during the integration time for a plurality of
pulses.
BRIEF DESCRIPTION OF THE DRAWINGS
Other features and advantages of the invention become apparent from
the following description of certain embodiments of the invention
given with reference to the accompanying drawings, in which:
FIG. 1 is a block diagram representing a method in accordance with
the invention of synthesizing a replica,
FIGS. 2a to 2c are block diagrams showing one way of determining a
calibration pulse in the method from FIG. 1,
FIG. 3 is a block diagram of radar in accordance with the invention
using the replica determined by the method shown in FIG. 1,
FIG. 4 is a diagram of a template used in the method of the
invention,
FIG. 4a shows one example of modulation of the transmitted
signal,
FIG. 4b is a graph showing the impulse response obtained with a
conventional weighted replica, and
FIG. 4c is a graph showing an impulse response obtained with a
replica in accordance with the invention.
MORE DETAILED DESCRIPTION
The frequency reference R(f) which will be used in the compression
filter and which is characterized by a complex function of the
frequency f is synthesized by the method shown in FIG. 1.
To this end, the starting point (block 10) is a conventional
(possibly weighted) sinc function impulse response, for example,
which corresponds to a transmitted signal without distortion whose
frequency varies as a linear function of time, i.e. an ideal
"chirp". This conventional impulse response, which is a function of
time, is a digital signal.
Next (block 12), a template is applied to the conventional impulse
response with the aim of attenuating the secondary, far and diffuse
lobes in accordance with the accuracy required of the radar.
FIG. 4 shows an illustrative example of a template indicating the
objective with an impulse response whose main lobe is that of the
ideal impulse response (RI) in terms of sinc(t) (sinc function) and
whose secondary lobes are equal to those in terms of sinc(t) less
20 dB. The template therefore has a rectangular central part 60 and
side parts 62 and 64 which are 20 dB below the central part.
Applying the template (block 14) yields the required impulse
response RI.sub.obj (t), that is to say:
Because the replica is a function of frequency, the Fourier
transform of the function RI.sub.obj (t) is taken, as shown by
block 16. In this example a fast Fourier transform (FFT) is
used.
The function RI.sub.obj (f) when the Fourier transform has been
taken (block 18), which is a spectrum, is divided by the spectrum
S.sub.cal (f) of a calibration pulse (block 20).
The calibration pulse spectrum S.sub.cal (f) is determined (block
24) by taking the fast Fourier transform (block 23) of a measured
and/or estimated calibration pulse (block 22).
One example of calibration pulse measurement is described with
reference to FIGS. 2a to 2c.
Each of FIGS. 2a, 2b and 2c shows the main parts of the instrument,
namely a radar platform 70, an antenna module 72 and a RF signal
distribution (beam-forming) network 74.
The radar platform 70 comprises a digital module 76 which generates
the transmitted signal and samples it on reception and an RF module
78 which effects the frequency change and amplification. These two
modules are used for transmission and reception. For transmission
module 76 supplies an intermediate frequency signal to module 78
and for reception module 78 supplies an intermediate frequency
signal to module 76.
The active antenna 72 comprises a section 80 of active transmit and
receive modules and a set of radiating elements 82. Module 74
samples or injects RF signals into the antenna during
calibration.
A process made up of three steps is used to determine the
calibration pulse in an instrument of the above kind.
During the first step, shown in FIG. 2a, the signal transmitted by
the radar passes only through modules 76 and 78, in both
directions. A signal S.sub.D (f) having the following value is
therefore obtained at the receive output 84:
In the above equation S1(f) is the contribution to the distortion
of modules 76 and 78 in the transmit channel and S2(f) is the
contribution of modules 76 and 78 in the receive channel.
The transmit channel of the instrument is calibrated during a
second step shown in FIG. 2b. To this end the transmitted signal
passes through modules 76 and 78, the components of section 80
concerned with transmission, the network 74 and then modules 78 and
76 again.
In this case a signal SE(f) having the following value is obtained
at output 84:
In the above equation S1(f) and S2(f) have the same meaning as in
FIG. 2a described above and S.sub.E ant(f) corresponds to the
distortion introduced on transmission by the antenna 80 and the
network 74.
During the third step, shown in FIG. 2c, the transmitted signal
passes first through modules 76 and 78, then network 74, then the
components of the section 82 dedicated to reception and finally
modules 78 and 76 to output 84.
In this case a signal S.sub.R (f) satisfying the following equation
is obtained at output 84:
In the above equation S.sub.R ant(f) corresponds to the distortion
introduced in this receive mode by network 74 and by the components
of section 80 concerned with reception.
The calibration pulse S.sub.cal (f) then has the value:
##EQU1##
Thus it can be seen that the calibration pulse allows for the
distortion introduced on transmission and reception by modules 76
and 78, the antenna module and network 74.
However, this calibration is not perfect because it does not allow
for all distortion, in particular distortion that can be introduced
downstream of the components of section 80. However, experience
shows that this type of calibration gives satisfactory results.
In the case of radar in which successive coherent pulses are
compressed (for example synthetic aperture radar), the calibration
or estimation operation consists in measuring the variation in
amplitude and phase of the received signal pulse by pulse. This
variation is multiplied by a signal S(t) which is the ideal signal
sampled at the pulse frequency (for example a "chirp" with a
bandwidth equal to the Doppler bandwidth for synthetic aperature
radar). The replica is then synthesized as described with reference
to FIG. 1.
The synthesized replica R(f) is used in the conventional way to
obtain the response of the radar, as shown in FIG. 3 (block
20):
The active signal is first digitized. The digitized active signal
is represented by the function S(t) (block 30). The fast Fourier
transform of signal S(t) is then taken (block 32) to obtain signal
S(f) to be compressed, which is a spectrum (block 34).
To compress the signal, the spectrum S(f) is modified by the
synthesized replica R(f) (block 36). This yields the spectrum:
RI(f)=R(f).times.S(f)
The inverse fast Fourier transform of the compressed signal is then
taken (block 38) to obtain the impulse response of the radar in the
time domain (block 40).
Results obtained with pulse compression radar in accordance with
the invention where the transmitted signal corresponds to a "chirp"
with two discontinuities as shown in FIG. 4a will now be described.
The FIG. 4a diagram represents the variation over a frequency band
2B and as a function of time t of the frequency modulation of the
signal transmitted during a time interval T.
The diagram 90 includes a first segment 92 between times 0 and .pi.
(.pi. is small compared to the duration T of the transmitted
pulse), a second segment 94 between times .tau. and T-.tau. having
a slope less than the slope of the straight line segment 92, and
finally a third segment 96 between times T-.tau. and T which is a
straight line segment having the same slope as the first segment
92.
The FIG. 4b diagram shows the impulse response obtained with a
conventional replica equal to the conjugate of the transmitted
signal, perfect calibration and a Hamming weighting coefficient of
0.58.
FIG. 4c shows the impulse response obtained with a synthesized
replica in accordance with the invention, the template imposed on
the secondary lobes and on the diffuse lobes during synthesis being
in the same order of magnitude as the 0.58 Hamming weighting, as
shown in FIG. 4b.
In FIGS. 4b and 4c time in microseconds is plotted on the abscissa
axis, the origin corresponding to the peak of the main lobe
50.sub.1 or 50.sub.2.
The main lobe 50.sub.2 of the invention (FIG. 4c) is narrower than
the conventional main lobe 50.sub.1 (FIG. 4b) An improvement in
resolution in the order of 35% is obtained, the expansion factor
being reduced from 1.2 to 0.89. The resolution is the ratio between
the expansion factor and the bandwidth B of modulation of the
transmitted signal (FIG. 4a).
Also, with the invention, the signal to noise ratio is improved
approximately 1 dB.
For identical performance in terms of resolution, link balance and
secondary lobe levels for a point target, with the invention it is
possible to obtain a 35% reduction in the band transmitted, a 25%
reduction in transmitter power and in power consumption and
improved contrast of a point target relative to a diffuse target,
and therefore improved probability of detection.
The main lobe being 35% narrower, the required resolution is
obtained using a band B1 such that:
B.sub.1 =B.sub.conventional .times.0.65, because the resolution is
inversely proportional to the transmitted bandwidth.
Similarly, instead of exploiting the 1 dB improvement in the signal
to noise ratio obtained by the processing method, it is possible,
for the same performance in terms of signal to noise ratio (SNR),
to reduce the average power transmitted (and therefore power
consumption) by 25%, comparing radar using the principle of the
invention to radar with conventional processing.
Also, as the main lobe is not attenuated, the contrast between the
point target (main lobe) and diffuse target (noise floor) is
increased commensurately, which improves detection performance.
The method of the invention applies to any type of pulse
compression or coherent radar and in particular to imaging radar,
rainfall radar, altimeters, wind diffusion meters, etc. Imaging
radar (SAR) is pulse compression radar in terms of range (the chirp
is in the transmitted pulse) and azimuth, i.e. along the speed
vector, in the coherent integration time corresponding to
traversing the antenna diagram.
Rainfall radar can use the pulse compression technique provided
that the technique adopted guarantees a very low level of the
secondary lobes, typically -60 dB, to prevent pollution of rainfall
echoes by ground echoes. Pulse modulation of the type shown in FIG.
4a is then employed.
Wind diffusion meters are also compression radars on the range axis
to improve the link balance.
Similarly, altimeters are compression radars on the range axis.
In all these radars optimization necessitates a compromise between
the impulse response parameters (resolution, level of secondary
lobes), radar parameters (bandwidth, pulse duration, signal to
noise ratio) and compression algorithms.
Table I below is a comparative table showing the performance
obtained from radar conforming to the invention (synthesized
replica) having in-band undulation type distortion and the
performance obtained with a weighted but uncalibrated replica and
with a calibrated conventional replica, to which a 0.58 Hamming
weighting coefficient is applied.
Compared to uncalibrated or calibrated replica pulse compression
radars, the level of the main peak is considerably improved, i.e.
the signal to noise ratio is improved, and resolution and ISLR also
show a clear improvement.
TABLE I ______________________________________ Non- Calibrated
calibrated "conventional" Synthesized RI parameter replica replica
replica ______________________________________ Main peak -4.9 dB
-4.7 dB 0 dB level Resolution 1.2 1.2 0.89 expansion factor ISLR
17.2 dB 19.6 dB 25.6 dB PSLR 19.8 dB 34.3 dB 34 dB SSLR 41.6 dB 43
dB 44.9 dB ______________________________________
In the above table, ISLR signifies "Integrated Side Lobe Ratio" and
conventionally corresponds to a ratio between the energy in the
central lobe (between the maximum and -3 dB) and the energy outside
the central lobe.
PSLR signifies "Peak Side Lobe Ratio" and corresponds to the
distance in dB between the maximum of the main lobe and the maximum
of a secondary lobe in a particular window around the main
peak.
SSLR signifies "Spurious Side Lobe Ratio" and corresponds to the
distance in dB between the maximum of the main lobe and the maximum
of a secondary lobe outside the aforementioned window.
A second example uses radar transmitting a conventional chirp, i.e.
linear frequency modulation of the transmitted signal with a band
B--20 MHz and a duration T 25 .mu.s, the sampling frequency being
1.1 B on the output side of an anti-aliasing filter.
Table II below compares radar with an impulse response obtained in
the conventional way and radar of the invention with a synthesized
impulse response.
TABLE II ______________________________________ "Conventional"
RI
"Synthesized" RI + RI parameter 0.58/0.65 Hamming 0.65 Hamming
______________________________________ Losses on main 1 dB/0.6 dB
0.6 dB peak (associated with weighting) Resolution 1.2/1.1 1.1
expansion factor PSLR 34.3 dB/26.8 dB 33 dB ISLR 19.7 dB/19 dB 19.9
dB SSLR 43.3 dB/38 dB 43.6 dB
______________________________________
In the conventional radar two Hamming weighting coefficients were
considered, one of 0.58 and the other of 0.65. In the case of the
invention, the impulse response was synthesized from an impulse
response in the time domain obtained by the conventional method
with a Hamming weighting coefficient of 0.65 and a 20 dB
attenuation template on the secondary lobes and with the band of
the replica limited to 1.1 times the band B of the transmitted
chirp.
Table II shows that the synthesized impulse response had secondary
lobes below 33 dB, i.e. substantially at the same level as those
obtained with conventional Hamming weighting and a coefficient of
0.58. The loss on the signal to noise ratio and the degraded
resolution were practically identical to those obtained with a
conventional method and a Hamming coefficient of 0.65.
Table II shows clearly that with a single coefficient the invention
combines the advantages obtained with conventional 0.58 and 0.65
weighting coefficients without their disadvantages.
* * * * *